The conditions for a pluralistic digital future: interoperability, transparency, and control over data
A reflection on the necessary conditions to be not only “users”, but also “citizens” of the digital spaces in which we live: owners of rights, owners of one’s own personal data, free to choose which services to use, which suppliers to turn to, and free to know what criteria determine what appears on our screens
In a piece dealing with the role of tech giants published in Valigia Blu earlier this year, Bruno Saetta highlighted the importance of interoperability between services, and argued that “in the absence of abuse” there is no need to “break” the tech giants. In dialogue with that article, I propose here additional considerations in favour of the urgency of addressing the issue of interoperability, referring also to technical aspects that make its introduction possible (though not simple), and highlighting how in some cases this may lead to the need to vertically break some of the great oligopolies of the Silicon Valley.
It is not just about bringing back competition in some market sectors, stimulating innovation, protecting privacy, creating space for different business models, or offering alternatives to consumers who, for example, want to leave Facebook without becoming digital hermits. It is also a question of allowing us – as individuals and citizens rather than just users – to make choices, to feel in control of the digital spaces in which we invest a significant part of our lives, and to introduce transparency and pluralism in a context in which decisive choices for our access to information are made without our knowledge and without any real possibility for us to understand or change them. As users, letting others make choices for us is often the easiest thing. As citizens, if these choices directly concern central aspects of the functioning of a democratic society, it is simply unacceptable.
These are wide, complex issues, partly determined by technological, economic, and legal processes. Let’s proceed step by step, starting from the discussion proposed by Bruno Saetta on the importance of interoperability .
Interoperability used to be the norm
Since telephones became widespread at mass level, it has always been possible to call other people simply by selecting the relevant telephone number; since various private players entered the mobile market, it has always been possible to call one operator from another. Historically, the same has always been true for what is the most consolidated method of communication on the Internet, the email: if I have a Hotmail account I can write to those who have a Gmail account, and not being able to do so would seem absurd.
In the user’s perspective, it is difficult to imagine arguments against interoperability between services. In the perspective of the company that offers services, above all from the moment in which it manages to conquer an important share of users, there are instead strong motivations to create a closed product (or ecosystem of products). To appreciate the contrast between the two approaches, it is enough to compare e-mail services and instant messaging services. In the case of email, I can freely choose the provider of my email account and the interface I prefer to use (web services, desktop applications such as Outlook and Thunderbird, different phone apps, etc.). On the contrary, if I have a messaging account with WhatsApp, I can only use the WhatsApp app to interact with other Whatsapp users. I cannot interact with Whatsapp users via an account obtained from another provider. From a Telegram account I cannot exchange messages with a user of Facebook Messenger, Signal, or Google Hangouts.
However, even in the context of instant messaging, this has not always been the case. For example Google Talk, the chat service that was available for years from the Gmail interface, was proudly interoperable. On the service description page, Google explicitly stated the importance of allowing users to use different interfaces (" We believe that users should be able to choose which program to use to connect to Google Talk"), and of allowing Google Talk users to interact with others who used a compatible protocol (XMPP). In a comparison with email services, Google stated that, “unfortunately”, the same freedom was not found in instant messaging services, and boasted interoperability as an important feature of the product:
“ The Google Talk network allows open interoperability with hundreds of other communications service providers through a service known as the ‘federation’. This means that a service user can interact with users of other services without the need to create new accounts or make new registrations”.
In 2013, however, Google abandoned Google Talk and introduced Google Hangouts, completely renouncing the concept of interoperability.
Advantages and disadvantages of interoperability
The first advantage for large companies that opt for closed systems is certainly that of “vendor lock-in”, a mechanism that makes it costly (often artificially so) to leave the service and move on to another. It is important to highlight that it is not necessarily about economic costs: most users “cannot afford” to abandon Whatsapp or Facebook for the main reason that they would find almost no one on alternative services and do not want to lose the contacts they already have (see the concept of “ network externalities"). Even if alternative services offered a better product (more privacy, less ads, or simply a more appealing interface), the cost of abandoning a dominant network would remain prohibitively high: an effective, privacy-friendly messaging interface is useless if there are no other users to communicate with.
It is fair to recognise, however, that interoperability can be costly from the point of view of a product developer. A classic problem with federated systems such as email is that of spam, although the problem is less significant in the case of messaging systems that only accept authentication through a single phone number, such as the ones mentioned above. Overall, for a large company it is much easier to choose new standards and impose them on its users than to negotiate the evolution of a protocol through dedicated working groups involving numerous partners. From this perspective, the email would also act as a negative example: it is based on a standard so old and implemented in so many different ways that it would be extremely difficult to update it, for example to make communication more private. This line of reasoning has determined for example the choice of Signal (an instant messaging programme particularly attentive to the issue of privacy) to refuse any form of interoperability, even if it is an open source product backed by donations.
These are well-founded arguments but, if generalised, also very questionable and questioned. In many fields, in fact, it is precisely shared standards and protocols that allow faster technological progress, or avoid situations of great inconvenience for users. The “vendor lock-in” also acts as a major brake on innovation: the main concern of the technology giants is not so much to create a better product, but rather to find a better way to squeeze profits by exploiting the user base that they have “caught”.
Interoperability in practice
Is it so unimaginable to think that users of similar services can interact with each other without needing to register on other services? Apparently not, and Facebook might just be the one to prove it: the American company has in fact announced its intention to integrate the instant messaging services of Facebook, Instagram, and Whatsapp (the technical details of the operation are not yet known). To date, Facebook Messenger, Whatsapp, and Skype already use the Signal protocol in whole or in part, without allowing federation between them.
Furthermore, various federated messaging services already exist and operate, among which the most promising currently seems to be Matrix , recently chosen for example to manage communications of France’s government apparatus . A new phone model to be released in 2019 will offer Matrix as a default option for messages and calls, and the company that offers it allows you to register on their own server. By its nature, this new server is interoperable with other existing Matrix servers or servers that even individual users, communities, or user cooperatives can in turn maintain: the protocol is shared, and therefore everyone can freely choose which server to register on and which programme to use to communicate. Just like email.
As for social networks, in January 2018 the consortium that defines Internet standards (W3C) recommended a new standard for federation between social networks,
ActivityPub , used by
hundreds of thousands users of
Mastodon and other services that support the standard, many of which are in development (among them
PeerTube focused on videos, or
Pixelfed on images). Also in this case it is possible to register on servers with different interfaces, rules, and principles, and to interact with those registered elsewhere: you can opt for generalist public servers, thematic, for a fee,
managed by a cooperative , or have a server for a single user, from which however they interact with all the others.
There is a lot of potential, and certainly also many problems, linked to the direct interaction between compatible services, especially when private data are involved. In these years, as a society, we have delegated to the tech giants so many choices, including the responsibility to decide what can be legitimately published in a shared space like social networks: in a decentralised system, each entity could reasonably set up different rules, for example allowing alternative approaches to managing the flow of contents that are shown to users, benefiting pluralism and freedom of expression, but also creating new problems that the technology giants currently solve for us, including security, moderation, and control of access to data.
This is a path that is by no means to be taken for granted, but which would offer new opportunities, including greater freedom of choice and effective competition and innovation mechanisms. As in other markets, in all likelihood some large suppliers would grab an important share of users but, unlike what we see today, new private, public, for-profit or non-profit, or cooperative actors could also find space. As presented in a piece by Niccolò Caranti published on our site, it is not unthinkable that Facebook could work on a governance basis similar to that of a non-profit foundation like the one that is at the base of Wikipedia.
The challenges are many, but if the alternative is further centralisation of services in the hands of a few companies that limit the power of choice of citizens at various levels, for example by making it impossible for them to know why they are faced with a certain piece of news instead of another – these are questions for which we must try to find solutions and promote them, without waiting for a solution to come from Silicon Valley’s “venture capital” investors who have created the current context and benefit from it.
“Break” the technology giants
As discussed, interoperability is fundamental to having pluralism in the information and communication technology ecosystem in which many of us now spend an important part of our time, and which increasingly serves as a privileged source of information. In short, these are all but neutral questions from the point of view of a democratic system.
However, interoperability alone would not be sufficient to radically change the functioning of sectors such as social media (in part linked to the distribution of news and advertising), instant messaging, or others, in a context in which there are already pre-existing dominant actors. How to deal with a context in which few companies have “too much power” in a given sector? In the US and European Union context, it is not enough to show that a company is dominant in a sector to resort to anti-monopoly legislation, it is also necessary to find an “abuse of dominant position”. The criteria included in the EU treaties do not allow to give a clear answer to the question, partly because one of the key elements of the current framework is price: in a context in which the services of Google, Facebook, Twitter, etc. are offered for free, it is difficult to argue that the price is too high.
This is true, however, only if it is deemed that the price to be paid must necessarily be paid in cash. In the current context, in essence, the user “pays” for the service by transferring personal data. If this premise is accepted, it is difficult to argue that technology giants are not abusing their dominant position to “extort” more personal data than users would actually be willing to give up if they could freely choose in a more competitive context. From a legal point of view, we are on slippery ground, from which we probably cannot get out of without legislative intervention. Given the growing role of data in the digital economy, and considering that with our data we “pay” access to all the “free” digital services that have become an important component of our daily lives, it therefore seems urgent to update the legislation on competition explicitly including the transfer of data as a valid element in determining the abuse of dominant position.
Beyond the legal issue, economic and technical issues remain about how these tech giants could possibly be “broken”. The simplest way is to separate company branches which, among other things, were originally separate, or at least block future acquisitions, in spite of the fact that competent authorities have in the past approved such mergers (for example by allowing Facebook to buy Whatsapp and Instagram). But if the term of comparison is that of public utilities, it may be more relevant to break up companies vertically, as was done for example in the case of telephone/Internet access and electricity. Without going into the details of how these systems work in practice and how they got to their present status, it is worth mentioning that, for example in the energy sector in Italy, the infrastructure is managed by a third party ( Terna ), which guarantees access to the network at conditions and prices regulated by an authority ( AEEG). In the context of telephone and Internet access, a rather similar logic has allowed the liberalisation of the “ last mile “, which is the mechanism that allows various private companies to compete without having to build their own infrastructures, let alone without having to convince users to break ties with all their contacts using other operators.
How could this logic be applied in the context of social media? The aspects to focus on would be primarily two: access to users’ personal data and access to customers, i.e. – lest we forget it – who pays for advertising.
Actually private data, or data shared for the common good
Is it possible to imagine that the personal data that we typically entrust to social networks are instead managed by a separate entity, with adequate safeguards, and access capabilities offered at the user’s discretion, but in principle not tied to any single company? To be clear, the data that I have given over to Facebook in these years would no longer be in Facebook’s hands, but would be managed by a third party and under my control: I could decide to delete them at will, or give access to someone who is not Facebook, under exactly the same conditions as Facebook has access to.
A variant of this model is supported for example by Tim Berners-Lee , known as the “founder of the web”. According to the model he supports, each user would have a “pod” to store their personal data, and would be able to decide who and under what conditions could access them. This approach brings new problems and challenges: as has emerged in recent years, users often guarantee access to their data without a second thought (which happens whenever you install an app on a smartphone, for example) and do not use secure login credentials. On the other hand, this is in fact the condition we already have today, only that it is companies like Facebook that have actual control over the “pod” with our data and decide who can access it and under what conditions. As extensively argued by Aral Balkan – an activist and a supporter of decentralised solutions with characteristics similar to those of the “pods” described above – these are data that are increasingly an integral part of who we are, and as such should be more fully protected.
The other side of the problem is that in the current context some of the data that we give to the tech giants of the Silicon Valley could be used for the common good; for example, the traffic data collected by Google Maps enrich Google, but only Google is in a position to decide whether and how to provide part of this data to public administrations to develop better services. An alternative model would leave the user the choice to keep their data private or to share them for the common good (this is essentially the motto of the Decode consortium of municipalities and researchers, financed by the European Union).
Breaking the big technology companies vertically would also mean separating what in recent years has been the primary source of revenue for many of them: advertising. As with personal data, it would therefore be desirable for advertising to be managed by an independent, regulated third party. Various companies could compete in offering their services to customers (those who pay for advertising), and offer them to service providers (for example, social platforms), or directly to end users. The end user could then realistically choose a different supplier, and decide whether to be exposed to advertising, whether to pay and not be exposed, and perhaps how to distribute (also considering content producers) the revenue generated by these mechanisms.
Such a system would simplify the process of making advertising fully transparent in the web age, an element that – as discussed in a previous article – is increasingly important to ensure the integrity of public elections and debates in democratic contexts.
In line with the arguments presented by Bruno Saetta in the aforementioned article , this piece supports the importance of interoperability between digital services to make “users” primarily “citizens”: owners of rights, owners of their personal data, free to choose which services to use, which suppliers to turn to, and what criteria determine what appears on their screens when they turn on their phone or computer. In contrast to Saetta’s conclusions, this article argues in favour of the need to break the tech giants, adapting approaches observed in the past in other important sectors of the economy and communication, also fully recognising the economic value of data among the elements which determine the abuse of a dominant position.
The creation of a plural system of digital services based on interoperability and the actual and direct possession and control of personal data and digital assets by rights holders is a path fraught with technical, legal, and economic obstacles, but necessary in democratic contexts in which access to information cannot be filtered through opaque monopolies. These are issues that cannot be resolved exclusively by national legislators, but both national governments and local authorities (see for example the experience of the municipality of Barcelona ) can contribute to laying the foundations and strengthening the infrastructures that make this path ever more realistic.
Beyond proposals on specific aspects such as that aimed at breaking the tech giants outlined (in general terms) by the candidate for the Democratic primaries in the US, Elizabeth Warren , the elaboration of these themes has not yet fully entered the political and programme debate. More structured proposals like the one published by Diem 25 can still offer a model for other political forces that could certainly get to different answers, but should at least partly ask the same questions. A public debate involving various political forces is urgent: the alternative is to let others decide for us the rules that shape the privately-owned public spaces to which a growing part of our lives - as a society and as private citizens - is tied.